How Claude Cowork handles this
Cowork suits the job when you would rather hand over a task than steer each step. You write the goal in a sentence: raise the score on the docs site, preview first. Cowork uses the connectors on your Claude account, so the xSpeed Hub connector has to be added in Claude before Cowork can see it. From there it lists your sites, picks docs, and calls optimize_site with dry_run on. The preview returns the plan and changes nothing.
The useful part is that the plan comes back to you as a finished piece of work. Cowork can tell you which settings it would apply, what level it chose, and which modules the site has from list_modules. In Manual mode it pauses for your Allow or Deny before the real pass. When you approve, the pass applies each setting one at a time, checks the page after each change, and undoes anything that breaks it. Allow up to two minutes per site, so this is a task you can leave running while you do something else.
The result is the usual three lists: applied, undone with the reason, and unfixable. Read the unfixable list as a to-do for the site owner, because caching cannot shrink page weight or fix hotlinked images. Then deal with the verify step. The Hub gives the agent verify_urls and tells it to look at the pages before it reports success, since the plugin cannot run JavaScript. Ask Cowork to check them, and open one or two yourself.
Set up Claude Cowork once
Already connected? Skip to the prompts. Alternatives and troubleshooting are on the Claude Cowork guide.
1Put your sites in xSpeed Hub
Sign in at app.xspeedcache.com with Google or email; the Hub is free and has no site cap. Then connect each WordPress site from its own dashboard: click Connect Hub in the xSpeed Cache top bar, then Connect via xSpeed Hub. Each site needs the free xSpeed Cache plugin.
2Add xSpeed Hub to your Claude account
Cowork does not keep its own connector list. In Claude, go to Customize > Connectors, click Add custom connector, name it xspeedhub, paste the URL, choose Sign in now and, if asked, Register automatically. The full steps, including the Team and Enterprise route, are on the Claude page: /agent/claude/.
https://app.xspeedcache.com/xspeed/mcp
3Connect, then start a new Cowork session
Click Connect on the connector and approve access on the xSpeed Hub page. Cowork loads the connectors enabled for your account when a session starts, so open a new Cowork session after you connect. You can also switch the connector on from the + menu in the chat box.
Customize > Connectors > xspeedhub > Connect
4Set tool permissions before the first task
Open the connector under Customize > Connectors and set each tool to Always allow, Needs approval or Blocked. A sensible start is Always allow for the read tools and Needs approval for purge_cache, toggle_cache, update_settings and the other writes. Then pick a Cowork mode; Manual is the one that shows you each request.
Tool permissions: get_cache_status = Always allow Tool permissions: purge_cache = Needs approval
Full Claude Cowork setup, sign-in options and FAQ
Before you send a prompt that changes something
Reads change nothing on your sites, though contact_support emails xSpeed support. Writes run once your connection allows them, so your client's approval prompt and a read-only connection (the token's Read-only everywhere switch, or a Viewer sign-in) are the gates that matter. Cowork has three modes. Manually approve (Manual) pauses and asks you to Allow or Deny each action. Automatically approve (Auto) keeps working and blocks what Claude judges to be unsafe. Skip all approvals (Skip) asks nothing and checks nothing. Each connector tool also has a permission, set under Customize > Connectors. In Manual mode, Always allow runs, Needs approval asks and Blocked is denied; in Skip mode only Blocked stops a call. xSpeed Hub has no confirmation step of its own, so under Skip, or for any tool set to Always allow, a write runs as soon as your connection allows writes. Auto relies on Claude's own judgement, which is not a lock.
What do I ask?
Three prompts written for Claude Cowork. More for this job are below.
Use xSpeed Hub to raise the PageSpeed score on the docs site. Preview the plan first and stop there so I can read it.
The plan looks fine. Apply it with xSpeed Hub, then check the verify URLs and write me a short summary of what changed and what is left.
Using xSpeed Hub, check whether the docs site has the Pro modules a higher target would need before we try one.
What happens, step by step
optimize_site is the Hub tool for "make my site faster". It measures the site, applies the recommended xSpeed settings one at a time, checks the page still renders after each change, undoes anything that breaks it, and reports what it applied, what it undid and what it could not fix. It reaches caching and asset delivery only. Page weight, images hosted on another domain and heavy video are outside it, and the report says so instead of claiming a win.
01
See where the site stands
The agent reads get_score_history for the last stored scores and list_modules for which modules exist on that site and what state they are in. Neither changes anything.
02
Preview with dry_run
optimize_site with dry_run: true returns the plan: each change and its tier, what would be skipped (already on, or needing a higher aggressiveness) and what it cannot fix. Nothing is applied. The tool description tells the agent to prefer this before the first real run on a site you have not optimized before.
03
Apply it
optimize_site without dry_run applies the plan. aggressiveness is safe (removals and server-side changes only), standard (the default) or aggressive, which includes settings known to break some themes and should be opted into deliberately. A run can take up to about 2 minutes per site.
04
Look at the site afterwards
When changes were applied, the result carries verify_urls and a note. The site checks its own HTML and reverts anything that breaks it, but it cannot run JavaScript, so verified: true does not prove the page works in a browser. The Hub tells the agent to open those URLs and check the page renders and the console is clean, or to tell you which URLs to check.
05
Reach for a number only if you named one
target_score (1 to 100, Pro sites only) turns one pass into repeated rounds toward that score, up to max_rounds, with 3 as the most per site. Each round is a real PageSpeed measurement and up to about 2 minutes. The result says why it stopped, and diminishing returns means what is left is outside what caching can reach.
06
Critical CSS on Pro sites
generate_critical_css generates above-the-fold CSS so the first screen can paint without waiting for the full stylesheets. It needs the Pro Critical CSS module on that site, and the site answers with an unknown-tool error without it. Pages are rebuilt with it on their next build, so the agent purges the cache afterwards.
Reference
| Main tool | optimize_site (write): measure, apply one setting at a time, verify, undo what breaks, re-measure |
|---|---|
| Preview | dry_run: true returns the plan and changes nothing |
| aggressiveness | safe, standard (default) or aggressive |
| Every site at once | site: "all" or a list of handles, one result per site, sites run one after another; the Hub's confirm-once instruction does not name optimize_site, so your client's prompt is the gate |
| Time | Up to about 2 minutes per site |
| target_score | 1 to 100; needs Pro on the site; without it the site runs one pass and says so |
| max_rounds | Up to 3 per site; only sent along with target_score; 12 rounds in total per call |
| Wide fan-out with a target | Each site gets fewer rounds, and the result carries a note saying so |
| Result fields | applied, skipped, reverted, unfixable, score, verified, rounds; verify_urls after a run that changed something |
| Already tuned site | Returns "everything is already on" plus what it cannot fix; that is an answer, not a failure |
| Critical CSS | generate_critical_css (write), Pro Critical CSS module required; purge_cache afterwards |
| Read-only connection | optimize_site and generate_critical_css are refused; dry_run is still a write call, so it is refused too. Applies to the connection token with Read-only everywhere on and to a Viewer sign-in, not to other OAuth sign-ins |
Rules worth keeping
- Run dry_run first. optimize_site changes live site settings as soon as your connection allows writes. The tool description tells the agent to preview first on a site that has not been optimized, but that is guidance, so your client's approval prompt and a read-only connection (the connection token with Read-only everywhere on, or a Viewer sign-in) are the real gates.
- Keep aggressive for a deliberate choice. It includes settings known to break some themes, and the agent should say so before using it.
- After a run that changed something, open the pages in verify_urls. The HTML checks cannot run JavaScript, so a page can pass them and still be broken in a browser. Do not call a run safe because verified is true.
- Relay the unfixable list. When a site is already tuned, the honest answer is that caching is done and the rest is page weight, hotlinked images or DOM size. The agent should not retry the tool or claim a gain it did not make.
- Only ask for target_score when you want the site driven to a number. Each round costs a real PageSpeed measurement, and a plain "make it faster" is one pass.
Good to know with Claude Cowork
Cowork has three modes. Manual pauses for your Allow or Deny on each action, Auto blocks what Claude judges unsafe, and Skip asks nothing and checks nothing. Each connector tool also has its own setting under Customize, Connectors: Always allow, Needs approval or Blocked. Use Manual and keep optimize_site on Needs approval. Under Skip, or with the tool on Always allow, the real pass starts as soon as your connection allows writes, because the Hub has no confirmation step of its own. Auto is a model judging its own agent, not a lock. A Viewer member's sign-in is always read-only and refuses optimize_site entirely, including the dry run, because the Hub counts it as a write tool. The Hub's Read-only everywhere switch covers only the connection token, not a Cowork OAuth sign-in.
More prompts for this job
They work in any client connected to xSpeed Hub.
Ask xSpeed Hub to preview what optimize_site would change on shop with a safe pass. Do not apply anything.
Use xSpeed Hub to optimize the blog site with the standard pass, then give me the pages to check by hand.
With xSpeed Hub, get docs to a mobile PageSpeed score of 90 if you can, and tell me why it stopped if it does not get there.
Use xSpeed Hub to show me what optimize_site would do on every site, as a dry run, and list which sites have the most left to change.
Use xSpeed Hub to generate critical CSS for shop, then purge the cache so the pages are rebuilt with it.
With xSpeed Hub, optimize shop, then compare the new PageSpeed score with the last stored one.
Using xSpeed Hub, what could optimize_site not fix on the shop site?
Frequently asked questions
Keep going
Raise the PageSpeed score with other agents
More with Claude Cowork
Documentation
- How to generate Critical CSS
- Performance recommendations
- How to run a speed test
- Free vs Pro: what is included
- How to write prompts for xSpeed Hub
- Claude Cowork + xSpeed
- Every AI agent that works with xSpeed
From the blog